Session 11: IO of Healthcare and Credit Markets

We evaluate the FDIC’s costs for resolving at-risk banks, as defined in Jiang et al., 2023 using the model developed in Allen et al., 2023 to determine bidder valuations of at-risk banks and simulate bidding. We show that resolution would cost the FDIC over $200 billion, if it maintained its bidder size and health restrictions, well in excess of the amount in its fund. Costs could be lowered if the FDIC relaxed these criteria and/or peeled off parts of the troubled banks. Our results illustrate that, during times of crisis, resolution costs can spiral because the set of unconstrained healthy bidders dries up.

Credit card lenders individualize contracts primarily through risk-based credit limits rather than interest rates. To understand lenders’ credit limit choices, I use novel statement-level data on the near-universe of UK credit cards active between 2010–2015 to estimate a structural model of the credit card market. The model explains differences in lenders’ credit limit distributions through a screening technology that provides lenders with a noisy signal of customers’ risk. I identify model parameters using a novel cost shock that results from the April 2011 case in the England and Wales High Court concerning the mis-selling of payment protection insurance. I use the estimated model to evaluate a counterfactual scenario in which lenders can freely individualize interest rates and credit limits, which the existing environment precludes. As a result, interest rates and credit limits are individualized, and profits increase. Risk-based interest rate discrimination emerges, resulting in large reductions in consumer surplus for the riskiest individuals. I conclude with potential explanations for the puzzling absence of risk-based pricing in the UK credit card market.

In this paper, we construct a hospital-level panel dataset covering the entire US hospital industry over the period 1996-2018. Using detailed hospital-level expenditure and revenue data, we measure hospital-level markups and document industry-wide increasing markups of about 18 percent. In addition, we confirm hospital-level markup estimates documented in the literature that were obtained using alternative approaches. We then relate markups to merger activity across hospital markets, and we find a strong association between markups and mergers at the market level. In particular, we find that mergers on average can explain about 3 of the 18 percent increase in markups. Our results indicate that market-level markups do not increase due to reallocation of market shares to higher markup hospitals, but rather by an increase of markups at hospitals involved in mergers. We confirm that hospital system-wide mergers are the main factor driving the estimated merger-markup effect.

This paper examines the impacts of health care provider exits on patient outcomes and subsequent reallocation. Using administrative data on the universe of nursing home patients, I estimate the mortality effects of 1,109 nursing home closures on incumbent residents with a matched difference-in-differences approach. I find that displaced residents face a short-run 15.7% relative increase in their mortality risk. Yet this increase is offset by long-run survival improvements, so the cumulative effect inclusive of the initial spike is a net decline in mortality risk. These gains are driven by patients reallocating to higher quality providers. I also find significant heterogeneity by local market conditions: the survival gains accrue only to patients in competitive nursing home markets, whereas residents in concentrated markets experience no survival improvements. I then develop and estimate a dynamic model of the nursing home industry with endogenous exit. Combining the model estimates with the mortality results, I examine the effects of counterfactual reimbursement policy experiments on nursing home closures and resident life expectancy. A universal 10% increase in the Medicaid rate decreases the frequency of closures but causes some low-quality providers to remain open in competitive areas. In contrast, targeted subsidies for facilities in areas with limited alternatives improve overall life expectancy by averting the costliest nursing home closures.

In the presence of adverse selection, there are two channels through which a merger affects welfare: a reduction in inefficient sorting and an increase in markups. Whether the benefit from improved sorting is sufficient to offset the cost of greater markups is an empirical question that depends on the merger. I capture this trade-off in a tractable discrete choice model and apply the model to potential mergers in the non-group insurance market regulated by the ACA. Even in the presence of transfers to address adverse selection, 13% of mergers lead to greater consumer surplus. In markets where the sorting distortion is greater than $7.5 per person, more than 1 out of 3 mergers improve consumer surplus. This highlights that antitrust enforcement should take the degree of adverse selection into account when assessing the potential harm from a merger.

The US health care system is rife with administrative burdens, but little is known about their causal effects on provider behavior. Using exogenous changes to the jurisdictions of Medicare Administrative Contractors, I show that the resulting increase in claim denials causes providers to adopt cost-saving technologies, bill more aggressively, and consolidate into larger practices. These endogenous responses by providers fully offset the mechanical reduction in Medicare spending we would expect from the increased claim denials. I explain this counterintuitive result using a model of firms’ investment in billing effort and technology. Estimates from this model show that investment costs amount to $89 billion per year and that increasing administrative burdens reduces providers’ profits by 4–6% while raising Medicare spending, making both providers and the government worse off. Counterfactual simulations indicate that increased administrative burdens would result in substantial reductions in health care spending if providers were unable to adjust investment, highlighting the short-run incentives insurers may have to raise administrative burdens.

Faced with rising health care costs, private health insurance plans in the United States increasingly require their enrollees to pay a share of health spending out of pocket, typically in the form of deductibles and copayments. We measure how these demand-side incentives affect the health care consumption of patients with chronic illnesses. However, for our measurement to be useful to inform policy design, we must overcome an identification challenge: the quantity response to a copayment change depends both on patients’ sensitivity to this price and the physician decision-makers’ knowledge about the relative prices patients face. We develop a novel moment inequality model that allows us to recover price sensitivity while imposing few assumptions on the decision-maker’s information about price. We first illustrate the value of our methodological approach in a simulated setting and then apply it in the context of physicians’ prescription drug treatment choices for type 2 diabetes. Using data from privately insured patients in Oregon, we find that physicians and patients are more price sensitive than alternative full-information empirical models might suggest. We also use our model to test physicians’ information about patient costs. The data suggest physicians use only aggregate price information in their choice, such as a drug’s average price in the prior year. We also identify heterogeneity in physicians’ information sets by their past experience with diabetes management.

This paper studies the role of commitment and competition in preventive care provision in health insurance. Limited consumer commitment reduces insurers' preventive investment in consumers' health because insurers cannot internalize all investment cost savings as consumers leave the insurer in the future. Competition amplifies underinvestment by increasing turnover while constraining market power. Exploiting a shift-share instrument for consumer turnover, I find turnover reduces insurers' per-member preventive investment and prevention utilization. I develop and estimate a model of dynamic insurer competition on preventive investment and premiums. Counterfactual analysis reveals preventive investment per enrollee triples when changing the market structure to a single private insurer. Average health risks decrease by 2.7\% to 6.5\%, or $167 to $406 per consumer. Consumer surplus changes are ambiguous and depend on the tradeoff between cost savings from alleviated commitment frictions and losses from market power. These results shed light on potential efficiency losses of fragmented payer markets due to investment externalities.

This paper studies how resale price maintenance (RPM), a vertical contract that allows upstream manufacturers to directly control downstream retail prices, affects firm competition and social welfare. RPM is controversial from a regulatory perspective since it can eliminate double markup but facilitate downstream price coordination. Exploiting a quasi-natural experiment of a government investigation against a leading pharmaceutical manufacturer on its fixed price RPM behavior and a novel pharmacy-level dataset, I document patterns supporting that RPM collapsed post-investigation. A difference-in-differences method shows that by adopting the RPM, the manufacturer decreases its products’ average retail prices by 5% while increasing the wholesale prices by 4%, therefore likely to be welfare-improving. Motivated by these findings, I estimate a structural model to decompose the welfare effects of RPM in this context and illustrate how the key market primitives (consumer substitution patterns and vertical bargaining power) interact with RPM. Counterfactual exercises show that consumer welfare would have been 77% higher had the price coordination incentives been turned off. In addition, RPM initiated by a more dominant upstream brand leads to anti-competitive outcomes.

Index funds are one of the most common ways investors access equity markets and are perceived to be a transparent and low-cost alternative to active investment management. Despite these purported virtues of index fund investing and the introduction of new products and competitors, many funds remain expensive and providers appear to exercise substantial market power. Why do index funds have market power? We develop a novel quantitative dynamic model of demand for and supply of index funds. In the model, investors are subject to inertia, search frictions, and have heterogeneous preferences. These frictions on the demand side create market power for index fund providers, which providers can further exploit by price discriminating and charging higher expense ratios to retail investors. Our results suggest that the average expense ratios paid by retail investors are roughly 70% higher as a result of search frictions and are 25% higher as a result of inertia compared to the friction-less baseline. Interestingly, there is an interaction between these frictions. Inertia becomes much more costly for investors when search frictions are reduced.

I study frictions that prevent banks and loan servicers from granting debt relief to struggling borrowers in the U.S. residential mortgage market. I explore how asymmetric information, transaction costs, and aid generosity associated with granting debt relief affect mortgage foreclosure outcomes. To disentangle these mechanisms, I introduce a structural model in which banks decide whether to offer debt relief to potentially distressed borrowers when processing relief is costly and borrowers hold private information about their financial well-being. Relative to full information, banks reduce the probability of granting relief to deter financially healthy borrowers from pretending to be distressed, leading to more foreclosures in equilibrium. I use my model to estimate the impact of the Federal Home Affordable Modification Program (HAMP) using the outcomes of mortgages that were originated before the 2008 financial crisis. I find that HAMP incentive payments offset bank costs enough to increase relief disbursement and to decrease realized foreclosures by 3%, or 200,000 properties nationally, over the decade from 2007 to 2016. Despite this, information frictions increased total foreclosures by 14%, or the equivalent of 1.1 million properties and $110 billion of lost value over the same time period. Finally, I find that the level of borrower relief prescribed under HAMP was insufficient for preventing 86% of foreclosures, highlighting the extent of borrower distress arising during 2008. Beyond the focus on malpractice in mortgage origination prior to the financial crisis, my findings illustrate how the design of debt relief and the behavior of financial intermediaries contributed to the widespread occurrence of foreclosure in the United States after the housing market collapse.

To address credit constraints in small-business lending markets, policymakers frequently use loan guarantees, which insure lenders against default. Guarantees affect loan prices by altering the effective marginal cost of lending but may create a moral hazard problem, weakening lenders’ information-acquisition incentives. I quantify these channels using data from the SBA 7(a) Loan Program. Guarantees benefit borrowers, on average, but redistribute surplus from low- to high-risk borrowers. Fixing government spending, an alternative policy with a 50% guarantee and a subsidy leads to a 1.0% increase in borrower surplus and a 0.1 percentage point (1.6%) decline in the program’s default rate.